btrfs: Enhance btrfs_trim_fs function to handle error better
diff mbox

Message ID 20171120055651.17892-1-wqu@suse.com
State New
Headers show

Commit Message

Qu Wenruo Nov. 20, 2017, 5:56 a.m. UTC
Function btrfs_trim_fs() doesn't handle errors in a consistent way, if
error happens when trimming existing block groups, it will skip the
remaining blocks and continue to trim unallocated space for each device.

And the return value will only reflect the final error from device
trimming.

This patch will fix such behavior by:

1) Recording first error from block group or device trimming
   So return value will also reflect any error found when trimming.
   Make developer more aware of the problem.

2) Outputting btrfs warning message for each trimming failure
   Any error for block group or device trimming will cause btrfs warning
   kernel message.

3) Continuing trimming if we can
   If we failed to trim one block group or device, we could still try
   next block group or device.

Such behavior can avoid confusion for case like failure to trim the
first block group and then only unallocated space is trimmed.

Reported-by: Chris Murphy <lists@colorremedies.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
---
 fs/btrfs/extent-tree.c | 59 ++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 43 insertions(+), 16 deletions(-)

Comments

David Sterba Nov. 20, 2017, 5:51 p.m. UTC | #1
On Mon, Nov 20, 2017 at 01:56:51PM +0800, Qu Wenruo wrote:
> Function btrfs_trim_fs() doesn't handle errors in a consistent way, if
> error happens when trimming existing block groups, it will skip the
> remaining blocks and continue to trim unallocated space for each device.
> 
> And the return value will only reflect the final error from device
> trimming.
> 
> This patch will fix such behavior by:
> 
> 1) Recording first error from block group or device trimming
>    So return value will also reflect any error found when trimming.
>    Make developer more aware of the problem.
> 
> 2) Outputting btrfs warning message for each trimming failure
>    Any error for block group or device trimming will cause btrfs warning
>    kernel message.

I think this could become too noisy, trimming failures are soft errors
IMO, so it should be enough to report all errors cumulatively
per-device.

> 3) Continuing trimming if we can
>    If we failed to trim one block group or device, we could still try
>    next block group or device.

Right, best-effort.

> Such behavior can avoid confusion for case like failure to trim the
> first block group and then only unallocated space is trimmed.
> 
> Reported-by: Chris Murphy <lists@colorremedies.com>
> Signed-off-by: Qu Wenruo <wqu@suse.com>
> ---
>  fs/btrfs/extent-tree.c | 59 ++++++++++++++++++++++++++++++++++++--------------
>  1 file changed, 43 insertions(+), 16 deletions(-)
> 
> diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
> index 309a109069f1..46d65ffb3bd1 100644
> --- a/fs/btrfs/extent-tree.c
> +++ b/fs/btrfs/extent-tree.c
> @@ -10948,6 +10948,16 @@ static int btrfs_trim_free_extents(struct btrfs_device *device,
>  	return ret;
>  }
>  
> +/*
> + * Trim the whole fs, by:
> + * 1) Trimming free space in each block group
> + * 2) Trimming unallocated space in each device
> + *
> + * Will try to continue trimming even if we failed to trim one block group or
> + * device.
> + * The return value will be the error return value of the first error.
> + * Or 0 if nothing wrong happened.
> + */
>  int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>  {
>  	struct btrfs_block_group_cache *cache = NULL;
> @@ -10958,6 +10968,8 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>  	u64 end;
>  	u64 trimmed = 0;
>  	u64 total_bytes = btrfs_super_total_bytes(fs_info->super_copy);
> +	int bg_ret = 0;
> +	int dev_ret = 0;
>  	int ret = 0;
>  
>  	/*
> @@ -10968,7 +10980,7 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>  	else
>  		cache = btrfs_lookup_block_group(fs_info, range->start);
>  
> -	while (cache) {
> +	for (; cache; cache = next_block_group(fs_info, cache)) {
>  		if (cache->key.objectid >= (range->start + range->len)) {
>  			btrfs_put_block_group(cache);
>  			break;
> @@ -10982,29 +10994,36 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>  			if (!block_group_cache_done(cache)) {
>  				ret = cache_block_group(cache, 0);
>  				if (ret) {
> -					btrfs_put_block_group(cache);
> -					break;
> +					btrfs_warn_rl(fs_info,
> +		"failed to cache block group %llu ret %d",

If this is meant for a developer as you say in the changelog, then
btrfs_debug is better.

> +						   cache->key.objectid, ret);
> +					if (!bg_ret)
> +						bg_ret = ret;
> +					continue;
>  				}
>  				ret = wait_block_group_cache_done(cache);
>  				if (ret) {
> -					btrfs_put_block_group(cache);
> -					break;
> +					btrfs_warn_rl(fs_info,
> +		"failed to wait cache for block group %llu ret %d",

The message wording is confusing, if this is another message for
developer, the function name can be printed.

> +						   cache->key.objectid, ret);
> +					if (!bg_ret)
> +						bg_ret = ret;
> +					continue;
>  				}
>  			}
> -			ret = btrfs_trim_block_group(cache,
> -						     &group_trimmed,
> -						     start,
> -						     end,
> -						     range->minlen);
> +			ret = btrfs_trim_block_group(cache, &group_trimmed,
> +						start, end, range->minlen);
>  
>  			trimmed += group_trimmed;
>  			if (ret) {
> -				btrfs_put_block_group(cache);
> -				break;
> +				btrfs_warn_rl(fs_info,
> +		"failed to trim block group %llu ret %d",
> +					   cache->key.objectid, ret);
> +				if (!bg_ret)
> +					bg_ret = ret;
> +				continue;
>  			}
>  		}
> -
> -		cache = next_block_group(fs_info, cache);
>  	}
>  
>  	mutex_lock(&fs_info->fs_devices->device_list_mutex);
> @@ -11012,15 +11031,23 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>  	list_for_each_entry(device, devices, dev_alloc_list) {
>  		ret = btrfs_trim_free_extents(device, range->minlen,
>  					      &group_trimmed);
> -		if (ret)
> +		if (ret) {
> +			btrfs_warn_rl(fs_info,
> +		"failed to trim unallocated space for devid %llu ret %d",
> +				      device->devid, ret);

So the idea is to print one message here, with devid, number of errors
and how many bytes were skipped.

> +			if (!dev_ret)
> +				dev_ret = ret;
>  			break;
> +		}
>  
>  		trimmed += group_trimmed;
>  	}
>  	mutex_unlock(&fs_info->fs_devices->device_list_mutex);
>  
>  	range->len = trimmed;
> -	return ret;
> +	if (bg_ret)
> +		return bg_ret;
> +	return dev_ret;
>  }
>  
>  /*
> -- 
> 2.15.0
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Qu Wenruo Nov. 21, 2017, 1:07 a.m. UTC | #2
On 2017年11月21日 01:51, David Sterba wrote:
> On Mon, Nov 20, 2017 at 01:56:51PM +0800, Qu Wenruo wrote:
>> Function btrfs_trim_fs() doesn't handle errors in a consistent way, if
>> error happens when trimming existing block groups, it will skip the
>> remaining blocks and continue to trim unallocated space for each device.
>>
>> And the return value will only reflect the final error from device
>> trimming.
>>
>> This patch will fix such behavior by:
>>
>> 1) Recording first error from block group or device trimming
>>    So return value will also reflect any error found when trimming.
>>    Make developer more aware of the problem.
>>
>> 2) Outputting btrfs warning message for each trimming failure
>>    Any error for block group or device trimming will cause btrfs warning
>>    kernel message.
> 
> I think this could become too noisy, trimming failures are soft errors
> IMO, so it should be enough to report all errors cumulatively
> per-device.

Although block group trimming errors are not that obvious to be reported
at per-device base.

> 
>> 3) Continuing trimming if we can
>>    If we failed to trim one block group or device, we could still try
>>    next block group or device.
> 
> Right, best-effort.
> 
>> Such behavior can avoid confusion for case like failure to trim the
>> first block group and then only unallocated space is trimmed.
>>
>> Reported-by: Chris Murphy <lists@colorremedies.com>
>> Signed-off-by: Qu Wenruo <wqu@suse.com>
>> ---
>>  fs/btrfs/extent-tree.c | 59 ++++++++++++++++++++++++++++++++++++--------------
>>  1 file changed, 43 insertions(+), 16 deletions(-)
>>
>> diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
>> index 309a109069f1..46d65ffb3bd1 100644
>> --- a/fs/btrfs/extent-tree.c
>> +++ b/fs/btrfs/extent-tree.c
>> @@ -10948,6 +10948,16 @@ static int btrfs_trim_free_extents(struct btrfs_device *device,
>>  	return ret;
>>  }
>>  
>> +/*
>> + * Trim the whole fs, by:
>> + * 1) Trimming free space in each block group
>> + * 2) Trimming unallocated space in each device
>> + *
>> + * Will try to continue trimming even if we failed to trim one block group or
>> + * device.
>> + * The return value will be the error return value of the first error.
>> + * Or 0 if nothing wrong happened.
>> + */
>>  int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>>  {
>>  	struct btrfs_block_group_cache *cache = NULL;
>> @@ -10958,6 +10968,8 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>>  	u64 end;
>>  	u64 trimmed = 0;
>>  	u64 total_bytes = btrfs_super_total_bytes(fs_info->super_copy);
>> +	int bg_ret = 0;
>> +	int dev_ret = 0;
>>  	int ret = 0;
>>  
>>  	/*
>> @@ -10968,7 +10980,7 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>>  	else
>>  		cache = btrfs_lookup_block_group(fs_info, range->start);
>>  
>> -	while (cache) {
>> +	for (; cache; cache = next_block_group(fs_info, cache)) {
>>  		if (cache->key.objectid >= (range->start + range->len)) {
>>  			btrfs_put_block_group(cache);
>>  			break;
>> @@ -10982,29 +10994,36 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>>  			if (!block_group_cache_done(cache)) {
>>  				ret = cache_block_group(cache, 0);
>>  				if (ret) {
>> -					btrfs_put_block_group(cache);
>> -					break;
>> +					btrfs_warn_rl(fs_info,
>> +		"failed to cache block group %llu ret %d",
> 
> If this is meant for a developer as you say in the changelog, then
> btrfs_debug is better.

Btrfs_debug looks good to me.

> 
>> +						   cache->key.objectid, ret);
>> +					if (!bg_ret)
>> +						bg_ret = ret;
>> +					continue;
>>  				}
>>  				ret = wait_block_group_cache_done(cache);
>>  				if (ret) {
>> -					btrfs_put_block_group(cache);
>> -					break;
>> +					btrfs_warn_rl(fs_info,
>> +		"failed to wait cache for block group %llu ret %d",
> 
> The message wording is confusing, if this is another message for
> developer, the function name can be printed.

OK.

> 
>> +						   cache->key.objectid, ret);
>> +					if (!bg_ret)
>> +						bg_ret = ret;
>> +					continue;
>>  				}
>>  			}
>> -			ret = btrfs_trim_block_group(cache,
>> -						     &group_trimmed,
>> -						     start,
>> -						     end,
>> -						     range->minlen);
>> +			ret = btrfs_trim_block_group(cache, &group_trimmed,
>> +						start, end, range->minlen);
>>  
>>  			trimmed += group_trimmed;
>>  			if (ret) {
>> -				btrfs_put_block_group(cache);
>> -				break;
>> +				btrfs_warn_rl(fs_info,
>> +		"failed to trim block group %llu ret %d",
>> +					   cache->key.objectid, ret);
>> +				if (!bg_ret)
>> +					bg_ret = ret;
>> +				continue;
>>  			}
>>  		}
>> -
>> -		cache = next_block_group(fs_info, cache);
>>  	}
>>  
>>  	mutex_lock(&fs_info->fs_devices->device_list_mutex);
>> @@ -11012,15 +11031,23 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
>>  	list_for_each_entry(device, devices, dev_alloc_list) {
>>  		ret = btrfs_trim_free_extents(device, range->minlen,
>>  					      &group_trimmed);
>> -		if (ret)
>> +		if (ret) {
>> +			btrfs_warn_rl(fs_info,
>> +		"failed to trim unallocated space for devid %llu ret %d",
>> +				      device->devid, ret);
> 
> So the idea is to print one message here, with devid, number of errors
> and how many bytes were skipped.

About number of errors, did you mean also accounting the errors found in
block group trimming?
In that case, it may not be related to all devices of a block group.

So this doesn't look appropriate to account block group errors into
device error.

Thanks,
Qu

> 
>> +			if (!dev_ret)
>> +				dev_ret = ret;
>>  			break;
>> +		}
>>  
>>  		trimmed += group_trimmed;
>>  	}
>>  	mutex_unlock(&fs_info->fs_devices->device_list_mutex);
>>  
>>  	range->len = trimmed;
>> -	return ret;
>> +	if (bg_ret)
>> +		return bg_ret;
>> +	return dev_ret;
>>  }
>>  
>>  /*
>> -- 
>> 2.15.0
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
David Sterba Nov. 21, 2017, 3:01 p.m. UTC | #3
On Tue, Nov 21, 2017 at 09:07:01AM +0800, Qu Wenruo wrote:
> On 2017年11月21日 01:51, David Sterba wrote:
> > On Mon, Nov 20, 2017 at 01:56:51PM +0800, Qu Wenruo wrote:
> >> Function btrfs_trim_fs() doesn't handle errors in a consistent way, if
> >> error happens when trimming existing block groups, it will skip the
> >> remaining blocks and continue to trim unallocated space for each device.
> >>
> >> And the return value will only reflect the final error from device
> >> trimming.
> >>
> >> This patch will fix such behavior by:
> >>
> >> 1) Recording first error from block group or device trimming
> >>    So return value will also reflect any error found when trimming.
> >>    Make developer more aware of the problem.
> >>
> >> 2) Outputting btrfs warning message for each trimming failure
> >>    Any error for block group or device trimming will cause btrfs warning
> >>    kernel message.
> > 
> > I think this could become too noisy, trimming failures are soft errors
> > IMO, so it should be enough to report all errors cumulatively
> > per-device.
> 
> Although block group trimming errors are not that obvious to be reported
> at per-device base.

The idea is to give an early warning that some device is not all ok, but
fstrim might be the wrong place to do such checks and reports anyway.

> >> @@ -11012,15 +11031,23 @@ int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
> >>  	list_for_each_entry(device, devices, dev_alloc_list) {
> >>  		ret = btrfs_trim_free_extents(device, range->minlen,
> >>  					      &group_trimmed);
> >> -		if (ret)
> >> +		if (ret) {
> >> +			btrfs_warn_rl(fs_info,
> >> +		"failed to trim unallocated space for devid %llu ret %d",
> >> +				      device->devid, ret);
> > 
> > So the idea is to print one message here, with devid, number of errors
> > and how many bytes were skipped.
> 
> About number of errors, did you mean also accounting the errors found in
> block group trimming?
> In that case, it may not be related to all devices of a block group.
> 
> So this doesn't look appropriate to account block group errors into
> device error.

Right, the more I think about that, the per-device reports make less
sense.

After another look to btrfs_trim_free_extents, there are some error
cases that are not fatal and mostly related to interrupting the whole
operation, so this needs to be reworked.
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch
diff mbox

diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 309a109069f1..46d65ffb3bd1 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -10948,6 +10948,16 @@  static int btrfs_trim_free_extents(struct btrfs_device *device,
 	return ret;
 }
 
+/*
+ * Trim the whole fs, by:
+ * 1) Trimming free space in each block group
+ * 2) Trimming unallocated space in each device
+ *
+ * Will try to continue trimming even if we failed to trim one block group or
+ * device.
+ * The return value will be the error return value of the first error.
+ * Or 0 if nothing wrong happened.
+ */
 int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
 {
 	struct btrfs_block_group_cache *cache = NULL;
@@ -10958,6 +10968,8 @@  int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
 	u64 end;
 	u64 trimmed = 0;
 	u64 total_bytes = btrfs_super_total_bytes(fs_info->super_copy);
+	int bg_ret = 0;
+	int dev_ret = 0;
 	int ret = 0;
 
 	/*
@@ -10968,7 +10980,7 @@  int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
 	else
 		cache = btrfs_lookup_block_group(fs_info, range->start);
 
-	while (cache) {
+	for (; cache; cache = next_block_group(fs_info, cache)) {
 		if (cache->key.objectid >= (range->start + range->len)) {
 			btrfs_put_block_group(cache);
 			break;
@@ -10982,29 +10994,36 @@  int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
 			if (!block_group_cache_done(cache)) {
 				ret = cache_block_group(cache, 0);
 				if (ret) {
-					btrfs_put_block_group(cache);
-					break;
+					btrfs_warn_rl(fs_info,
+		"failed to cache block group %llu ret %d",
+						   cache->key.objectid, ret);
+					if (!bg_ret)
+						bg_ret = ret;
+					continue;
 				}
 				ret = wait_block_group_cache_done(cache);
 				if (ret) {
-					btrfs_put_block_group(cache);
-					break;
+					btrfs_warn_rl(fs_info,
+		"failed to wait cache for block group %llu ret %d",
+						   cache->key.objectid, ret);
+					if (!bg_ret)
+						bg_ret = ret;
+					continue;
 				}
 			}
-			ret = btrfs_trim_block_group(cache,
-						     &group_trimmed,
-						     start,
-						     end,
-						     range->minlen);
+			ret = btrfs_trim_block_group(cache, &group_trimmed,
+						start, end, range->minlen);
 
 			trimmed += group_trimmed;
 			if (ret) {
-				btrfs_put_block_group(cache);
-				break;
+				btrfs_warn_rl(fs_info,
+		"failed to trim block group %llu ret %d",
+					   cache->key.objectid, ret);
+				if (!bg_ret)
+					bg_ret = ret;
+				continue;
 			}
 		}
-
-		cache = next_block_group(fs_info, cache);
 	}
 
 	mutex_lock(&fs_info->fs_devices->device_list_mutex);
@@ -11012,15 +11031,23 @@  int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
 	list_for_each_entry(device, devices, dev_alloc_list) {
 		ret = btrfs_trim_free_extents(device, range->minlen,
 					      &group_trimmed);
-		if (ret)
+		if (ret) {
+			btrfs_warn_rl(fs_info,
+		"failed to trim unallocated space for devid %llu ret %d",
+				      device->devid, ret);
+			if (!dev_ret)
+				dev_ret = ret;
 			break;
+		}
 
 		trimmed += group_trimmed;
 	}
 	mutex_unlock(&fs_info->fs_devices->device_list_mutex);
 
 	range->len = trimmed;
-	return ret;
+	if (bg_ret)
+		return bg_ret;
+	return dev_ret;
 }
 
 /*